Authors
Christian Szegedy, Wojciech Zaremba, Ilya Sutskever, Joan Bruna, Dumitru Erhan, Ian Goodfellow, Rob Fergus
Publication date
2013/12/21
Journal
arXiv preprint arXiv:1312.6199
Description
Abstract: Deep neural networks are highly expressive models that have recently achieved
state of the art performance on speech and visual recognition tasks. While their
expressiveness is the reason they succeed, it also causes them to learn uninterpretable
solutions that could have counter-intuitive properties. In this paper we report two such
properties.
Total citations
2014201520162310584
Scholar articles
C Szegedy, W Zaremba, I Sutskever, J Bruna, D Erhan… - arXiv preprint arXiv:1312.6199, 2013